Reviewed-by: Matt Turner
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
If the ?: operator's condition is a constant value, and both branches
were pure expressions, we can just make the resulting value one or the
other.
Previously, we only did this if op[1] and op[2] were also constant
values - but there's no actual reason for that restriction.
No changes in shader-d
This adds a parent_instr field similar to the one for ssa_def. The
difference here is that the parent_instr field on a nir_register can be
NULL if the register does not have a unique definition or if that
definition does not dominate all its uses. We set this field in the
out-of-SSA pass so that
Here's take two of the fast mutex path. This version has a check for
linux/futex.h and it split into two patches. First patch introduces
full_mtx_t and updates callers. Second patch adds the inline mtx_t.
Kristian Høgsberg (2):
mesa: Split full featured mutex into its own type
mesa: Add new
Most of the mtx_t uses in mesa only use lock/unlock, but a couple of
places we use a condition variable and a recursive mutex. This patch
introduces a new full_mtx_t for the cases where we need a more featureful
mutex and changes mtx_t to only support lock/unlock.
Signed-off-by: Kristian Høgsberg
While modern pthread mutexes are very fast, they still incur a call to an
external DSO and overhead of the generality and features of pthread mutexes.
Most mutexes in mesa only needs lock/unlock, and the idea here is that we can
inline the atomic operation and make the fast case just two intruction
On Thu, Jan 29, 2015 at 6:03 PM, Chad Versace wrote:
> On 01/28/2015 01:45 PM, jim.br...@linux.intel.com wrote:
>> From: Jim Bride
>>
>> Add a EGLBoolean parameter unbound_only to _eglReleaseDisplayResources() to
>> allow that function to skip releasing contexts or surfaces that are currently
>>
Again, I think this would be more readable if the refactor was in its own
commit with a little more explanation what's going on. Then add DSA.
On Wed, Jan 21, 2015 at 5:40 PM, Laura Ekstrand
wrote:
> ---
> src/mapi/glapi/gen/ARB_direct_state_access.xml | 14 ++
> src/mesa/main/bufferobj.c
On 30.01.2015 03:52, Neil Roberts wrote:
> This limits the style changes to modes inherited from prog-mode. The
> main reason to do this is to avoid setting fill-column for people
> using Emacs to edit commit messages because 78 characters is too many
> to make it wrap properly in git log. Note tha
On Wed, Jan 21, 2015 at 5:40 PM, Laura Ekstrand
wrote:
> ---
> src/mapi/glapi/gen/ARB_direct_state_access.xml | 18 +++
> src/mesa/main/bufferobj.c | 156
> ++---
> src/mesa/main/bufferobj.h | 34 --
> src/mesa/main/tests/dispat
It might be good to say in the commit message "update glext.h to match
revision X from Khronos" and then put the comment about DSA functions
as a second line in the commit message.
--Jason
On Fri, Jan 23, 2015 at 2:23 PM, Laura Ekstrand
wrote:
> I checked, and all of the currently implement
On 01/28/2015 01:45 PM, jim.br...@linux.intel.com wrote:
> From: Jim Bride
>
> Add a EGLBoolean parameter unbound_only to _eglReleaseDisplayResources() to
> allow that function to skip releasing contexts or surfaces that are currently
> bound to a display. Setting unbound_only to EGL_FALSE prese
On 01/28/2015 01:45 PM, jim.br...@linux.intel.com wrote:
> From: Jim Bride
>
> In situations where a DRI2 driver is in use and a call to eglTerminate()
> is followed by a call to eglMakeCurrent() the mandatory call to glFlush()
> inside of dri2_make_current() would cause a segmentation fault beca
On 01/29/2015 12:00 AM, Kenneth Graunke wrote:
> Paul originally had to reverse engineer these formulas based on the
> description about how the sampler works. The description here is not
> the easiest to follow - especially given that it's from the Sandybridge
> era, when the hardware only did 4x
On Thu, Jan 29, 2015 at 6:06 PM, Brian Paul wrote:
> On 01/29/2015 02:20 PM, Roland Scheidegger wrote:
>>
>> Hi,
>>
>> the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
>> that indirections can only be done through the address reg and has quite
>> some extra code to deal with
On Thursday, January 29, 2015 17:00:08 Kenneth Graunke wrote:
> On Thursday, January 29, 2015 12:50:20 PM Jason Ekstrand wrote:
> > ---
> > src/glsl/nir/nir_algebraic.py | 20 +---
> > src/glsl/nir/nir_opt_algebraic.py | 12 +---
> > 2 files changed, 26 insertions(+), 6
On Thu, Jan 29, 2015 at 5:00 PM, Kenneth Graunke
wrote:
> On Thursday, January 29, 2015 12:50:20 PM Jason Ekstrand wrote:
> > ---
> > src/glsl/nir/nir_algebraic.py | 20 +---
> > src/glsl/nir/nir_opt_algebraic.py | 12 +---
> > 2 files changed, 26 insertions(+), 6 del
On Thursday, January 29, 2015 02:34:18 PM Jason Ekstrand wrote:
> total instructions in shared programs: 5998190 -> 5997603 (-0.01%)
> instructions in affected programs: 54276 -> 53689 (-1.08%)
> helped:293
> ---
> src/glsl/nir/nir_opt_algebraic.py | 30
On Thu, Jan 29, 2015 at 4:13 PM, Kenneth Graunke
wrote:
> On Thursday, January 29, 2015 12:50:16 PM Jason Ekstrand wrote:
> > This allows us to indicate a concept of an invalid type.
> > ---
> > src/glsl/nir/nir.h | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/src/glsl/nir/nir.h
On Thursday, January 29, 2015 12:50:20 PM Jason Ekstrand wrote:
> ---
> src/glsl/nir/nir_algebraic.py | 20 +---
> src/glsl/nir/nir_opt_algebraic.py | 12 +---
> 2 files changed, 26 insertions(+), 6 deletions(-)
>
> diff --git a/src/glsl/nir/nir_algebraic.py b/src/glsl
On Thursday, January 29, 2015 12:50:16 PM Jason Ekstrand wrote:
> This allows us to indicate a concept of an invalid type.
> ---
> src/glsl/nir/nir.h | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
> index 980fdd0..f2050de 100644
> --- a/src/glsl/n
Ping.
I'll handle the cherry-pick to 10.4 since a small code change is needed.
-Brian
On 01/27/2015 08:06 PM, Brian Paul wrote:
The _mesa_dlist_alloc() function is only guaranteed to return a pointer
with 4-byte alignment. On 64-bit systems which don't support unaligned
loads (e.g. SPARC or M
On 29/01/15 04:05, Michel Dänzer wrote:
> On 29.01.2015 00:37, Emil Velikov wrote:
>>
>> February 20th 2015 - Release candidate 3
>> February 20th 2015 - Release candidate 4
>
> I assume you meant February 27th for rc4. :)
>
Sigh... will I be able to write a statement without ever making silly
ty
On 01/29/2015 02:20 PM, Roland Scheidegger wrote:
Hi,
the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
that indirections can only be done through the address reg and has quite
some extra code to deal with this. Even though hardware and apis which
worked like that are defini
On Thu, Jan 29, 2015 at 5:31 PM, Rob Clark wrote:
> On Thu, Jan 29, 2015 at 4:20 PM, Roland Scheidegger
> wrote:
>> Hi,
>>
>> the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
>> that indirections can only be done through the address reg and has quite
>> some extra code to d
On 29/01/15 17:14, Kristian Høgsberg wrote:
> On Thu, Jan 29, 2015 at 6:36 AM, Emil Velikov
> wrote:
>> On 28/01/15 05:08, Kristian Høgsberg wrote:
>>> While modern pthread mutexes are very fast, they still incur a call to an
>>> external DSO and overhead of the generality and features of pthread
total instructions in shared programs: 5998190 -> 5997603 (-0.01%)
instructions in affected programs: 54276 -> 53689 (-1.08%)
helped:293
---
src/glsl/nir/nir_opt_algebraic.py | 30 --
1 file changed, 28 insertions(+), 2 deletions(-)
On Thu, Jan 29, 2015 at 4:20 PM, Roland Scheidegger wrote:
> Hi,
>
> the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
> that indirections can only be done through the address reg and has quite
> some extra code to deal with this. Even though hardware and apis which
> worked l
On Thursday, January 29, 2015 05:21:03 PM Neil Roberts wrote:
> According to the SkyLake bspec the 3DSTATE_CONSTANT_* commands only take
> effect on the next corresponding 3DSTATE_BINDING_TABLE_POINTER_* command. This
> patch just makes it set the BRW_NEW_SURFACES state when uploading the push
> co
On Thursday, January 29, 2015 11:14:27 PM Martin Peres wrote:
> Signed-off-by: Martin Peres
> ---
> src/mesa/drivers/dri/i915/intel_screen.c | 1 +
> src/mesa/drivers/dri/i965/brw_context.c | 3 +++
> src/mesa/drivers/dri/i965/intel_screen.c | 1 +
> 3 files changed, 5 insertions(+)
>
> diff --
On Thursday, January 29, 2015 11:14:26 PM Martin Peres wrote:
> While trying to understand a GLSL pass, curro and I tried running Unigine
> Tropics and the GLSL compilers would not compile the shaders.
>
> The reason is due to the fact that the shaders do not specify the needed GLSL
> version but
Roland Scheidegger writes:
> Hi,
>
> the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
> that indirections can only be done through the address reg and has quite
> some extra code to deal with this. Even though hardware and apis which
> worked like that are definitely old by
On Thu, Jan 29, 2015 at 12:50 PM, Jason Ekstrand
wrote:
> total instructions in shared programs: 8026469 -> 8026343 (-0.00%)
> instructions in affected programs: 28541 -> 28415 (-0.44%)
> helped:101
> HURT: 76
>
These numbers a
On Thu, Jan 29, 2015 at 12:50 PM, Jason Ekstrand
wrote:
> total instructions in shared programs: 8026509 -> 8026469 (-0.00%)
> instructions in affected programs: 6338 -> 6298 (-0.63%)
> helped:11
>
Ignore those numbers. they're bogus. Here are the real ones:
Now that we can scalarize with NIR, there's no need for all this code
anymore. Let's get rid of it and just do scalar operations.
v2: run copy prop before lowering phi nodes
v3: Get rid of the "emit(...)->saturate = foo" pattern
total instructions in shared programs: 5998321 -> 5974070 (-0.40%)
v2 Jason Ekstrand :
- Add better comments
- Use nir_ssa_dest_init and nir_src_for_ssa more places
- Fix some void * casts
v3 Jason Ekstrand :
- Rework the way we determine whether or not to sccalarize a phi node to
make the recursion non-bogus
- Treat load_const instructions as scalarizabl
Originally, get_alu_src was supposed to handle resolving swizzles and
things like that. However, now that basically every instruction we have
only takes scalar sources, we don't really need it anymore. The only case
where it's still marginally useful is for the mov and vecN operations that
are le
Hi,
for Gallium Nine we use ARR and UARL.
Do all cards supporting integers support indexing with an integer register ?
If yes we could remove our UARL usage.
As for ARR, if all cards are ok with normal registers, we could use a
normal one
too, as long as we can get the same rounding behaviour
Hi,
the address reg in tgsi is quite a nuisance. glsl-to-tgsi code assumes
that indirections can only be done through the address reg and has quite
some extra code to deal with this. Even though hardware and apis which
worked like that are definitely old by now.
Thus, I'm proposing the address reg
While trying to understand a GLSL pass, curro and I tried running Unigine
Tropics and the GLSL compilers would not compile the shaders.
The reason is due to the fact that the shaders do not specify the needed GLSL
version but also use in the same shader keywords that could never co-exist
because o
Signed-off-by: Martin Peres
---
src/mesa/drivers/dri/i915/intel_screen.c | 1 +
src/mesa/drivers/dri/i965/brw_context.c | 3 +++
src/mesa/drivers/dri/i965/intel_screen.c | 1 +
3 files changed, 5 insertions(+)
diff --git a/src/mesa/drivers/dri/i915/intel_screen.c
b/src/mesa/drivers/dri/i915/in
From: Rob Clark
Enable the 'sphinx.ext.graphviz' extension, and add in a section for
driver specific docs, with freedreno compiler docs beneath. The
goal is for more complete compiler docs, and hopefully some docs about
other parts of the driver (such as how tiling works, etc).
Note that there
On Wed, Jan 28, 2015 at 11:20 PM, Eric Anholt wrote:
> Jason Ekstrand writes:
>
> > I've got a patch I cooked up today that generalizes this a bit. I
> haven't
> > sent it yet because it lacks shader-db statistics. That said, I'm fine
> > with merging this in the mean time.
>
> Not sure what g
---
src/glsl/nir/nir_algebraic.py | 7 +++
1 file changed, 7 insertions(+)
diff --git a/src/glsl/nir/nir_algebraic.py b/src/glsl/nir/nir_algebraic.py
index f9b246d..75436f4 100644
--- a/src/glsl/nir/nir_algebraic.py
+++ b/src/glsl/nir/nir_algebraic.py
@@ -34,13 +34,18 @@ class VarSet(object):
total instructions in shared programs: 8026469 -> 8026343 (-0.00%)
instructions in affected programs: 28541 -> 28415 (-0.44%)
helped:101
HURT: 76
---
src/glsl/nir/nir_opt_algebraic.py | 26 ++
1 file chang
total instructions in shared programs: 8026509 -> 8026469 (-0.00%)
instructions in affected programs: 6338 -> 6298 (-0.63%)
helped:11
---
src/glsl/nir/nir_opt_algebraic.py | 5 +
1 file changed, 5 insertions(+)
diff --git a/src/glsl/nir/nir_opt_algebraic.py
There are some algebraic transformations that we want to do but only if
certain things are constants. For instance, we may want to replace
a * (b + c) with (a * b) + (a * c) as long as a and either b or c is constant.
While this generates more instructions, some of it will get constant
folded.
ni
---
src/glsl/nir/nir_algebraic.py | 20 +---
src/glsl/nir/nir_opt_algebraic.py | 12 +---
2 files changed, 26 insertions(+), 6 deletions(-)
diff --git a/src/glsl/nir/nir_algebraic.py b/src/glsl/nir/nir_algebraic.py
index 75436f4..6e7973d 100644
--- a/src/glsl/nir/nir_a
This allows us to indicate a concept of an invalid type.
---
src/glsl/nir/nir.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/glsl/nir/nir.h b/src/glsl/nir/nir.h
index 980fdd0..f2050de 100644
--- a/src/glsl/nir/nir.h
+++ b/src/glsl/nir/nir.h
@@ -576,6 +576,7 @@ void nir_alu_dest_copy(nir
This allows you to match on an unknown value but only if it is of a given
type. 90% of the uses of this are for matching only booleans, but adding
the generality of arbitrary types is no more complex.
nir_algebraic.py doesn't handle this yet but that's ok because the C
language will ensure that t
On 29/01/15 19:40, srol...@vmware.com wrote:
From: Roland Scheidegger
since the address reg holds integer values, ARL/ARR do an implicit float-to-int
conversion, so clarify that. Thus it is also incorrect to say that FLR really
does the same as ARL.
---
src/gallium/docs/source/tgsi.rst | 18 +
On Thu, Jan 29, 2015 at 2:40 PM, wrote:
> From: Roland Scheidegger
>
> since the address reg holds integer values, ARL/ARR do an implicit
> float-to-int
> conversion, so clarify that. Thus it is also incorrect to say that FLR really
> does the same as ARL.
Reviewed-by: Alex Deucher
> ---
>
From: Roland Scheidegger
since the address reg holds integer values, ARL/ARR do an implicit float-to-int
conversion, so clarify that. Thus it is also incorrect to say that FLR really
does the same as ARL.
---
src/gallium/docs/source/tgsi.rst | 18 --
1 file changed, 8 insertions(
On Thursday, January 29, 2015 04:20:59 PM Topi Pohjolainen wrote:
> Currently meta blit programs unconditionally write both color and
> depth values. This series introduces two flavours - with and without
> gl_FragDepth - allowing simpler programs for color only blits. Ken
> had observed that i965
I haven't worked in meta a lot, but I didn't see anything that didn't
look correct. The series is
Reviewed-by: Matt Turner
___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
Matt wanted shader-db numbers and here they are:
total instructions in shared programs: 5998287 -> 5974070 (-0.40%)
instructions in affected programs: 732041 -> 707824 (-3.31%)
helped:3135
HURT: 193
GAINED:
This limits the style changes to modes inherited from prog-mode. The
main reason to do this is to avoid setting fill-column for people
using Emacs to edit commit messages because 78 characters is too many
to make it wrap properly in git log. Note that makefile-mode also
inherits from prog-mode so t
On Thu, Jan 29, 2015 at 4:55 AM, Martin Peres
wrote:
> v2: review from Laura Ekstrand
> - use the refactored code to lookup the objects
> - improve some error messages
> - factor out the gl method name computation
> - better handle the spec differences between the DSA and non-DSA cases
> - quote
On Thu, Jan 29, 2015 at 4:55 AM, Martin Peres
wrote:
> v2: Review from Laura Ekstrand
> - give more helpful error messages
> - factor the lookup code for the xfb and objBuf
> - replace some already-existing tabs with spaces
> - add comments to explain the cases where xfb == 0 or buffer == 0
> - f
On Thu, Jan 29, 2015 at 6:21 AM, Topi Pohjolainen
wrote:
> diff --git a/src/mesa/drivers/common/meta.c b/src/mesa/drivers/common/meta.c
> index ec7564a..45035b1 100644
> --- a/src/mesa/drivers/common/meta.c
> +++ b/src/mesa/drivers/common/meta.c
> @@ -293,13 +293,14 @@ _mesa_meta_setup_blit_shader
Good idea.
Reviewed-by: Laura Ekstrand
On Thu, Jan 29, 2015 at 6:54 AM, Martin Peres
wrote:
> A student was wondering what was going on + I started working on it too.
>
> CC: Laura Ekstrand
> Signed-off-by: Martin Peres
> ---
> docs/GL3.txt | 11 ++-
> 1 file changed, 10 insertions(
According to the SkyLake bspec the 3DSTATE_CONSTANT_* commands only take
effect on the next corresponding 3DSTATE_BINDING_TABLE_POINTER_* command. This
patch just makes it set the BRW_NEW_SURFACES state when uploading the push
constants to ensure the binding tables will be updated.
This fixes the
On Thu, Jan 29, 2015 at 6:36 AM, Emil Velikov wrote:
> On 28/01/15 05:08, Kristian Høgsberg wrote:
>> While modern pthread mutexes are very fast, they still incur a call to an
>> external DSO and overhead of the generality and features of pthread mutexes.
>> Most mutexes in mesa only needs lock/un
A student was wondering what was going on + I started working on it too.
CC: Laura Ekstrand
Signed-off-by: Martin Peres
---
docs/GL3.txt | 11 ++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/docs/GL3.txt b/docs/GL3.txt
index 648f5ac..2dc1404 100644
--- a/docs/GL3.txt
+
On 28/01/15 05:08, Kristian Høgsberg wrote:
> While modern pthread mutexes are very fast, they still incur a call to an
> external DSO and overhead of the generality and features of pthread mutexes.
> Most mutexes in mesa only needs lock/unlock, and the idea here is that we can
> inline the atomic
Currently all blit programs are unconditionally compiled with
gl_FragDepth.
Signed-off-by: Topi Pohjolainen
---
src/mesa/drivers/common/meta.c | 3 ++-
src/mesa/drivers/common/meta.h | 1 +
src/mesa/drivers/common/meta_blit.c| 2 +-
src/mesa/drivers/co
Implementing an idea from Ken, on i965 the shader program for 2D
blits becomes significantly simpler.
Before:
pln(8) g6<1>Fg4<0,1,0>Fg2<8,8,1>F { align1 1Q compacted };
pln(8) g7<1>Fg4.4<0,1,0>F g2<8,8,1>F { align1 1Q compacted };
send(8) g2<1>UW g6<8,8,1>F
sampler
Signed-off-by: Topi Pohjolainen
---
src/mesa/drivers/common/meta.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/drivers/common/meta.c b/src/mesa/drivers/common/meta.c
index 45035b1..6ddec73 100644
--- a/src/mesa/drivers/common/meta.c
+++ b/src/mesa/drivers/common/m
Signed-off-by: Topi Pohjolainen
---
src/mesa/drivers/common/meta_generate_mipmap.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/mesa/drivers/common/meta_generate_mipmap.c
b/src/mesa/drivers/common/meta_generate_mipmap.c
index c38b46b..c1b6d3c 100644
--- a/src/mesa/driv
Currently meta blit programs unconditionally write both color and
depth values. This series introduces two flavours - with and without
gl_FragDepth - allowing simpler programs for color only blits. Ken
had observed that i965 behaves badly emitting five unnecessary moves
when the depth is included i
When color buffers alone are concerned the depth is not needed.
No regression on BDW where meta blit is used instead of blorp. I
also disabled blorp temporarily for fbo-blits on IVB and saw no
regressions there either.
I also compared several graphics benchmarks on BDW and saw neither
regressions
On 29/01/15 15:37, Fredrik Höglund wrote:
On Thursday 29 January 2015, Martin Peres wrote:
v2: Review from Laura Ekstrand
- generate the name of the gl method once
- shorten some lines to stay in the 78 chars limit
Signed-off-by: Martin Peres
---
src/mapi/glapi/gen/ARB_direct_state_access.x
On Thursday 29 January 2015, Martin Peres wrote:
> v2: Review from Laura Ekstrand
> - generate the name of the gl method once
> - shorten some lines to stay in the 78 chars limit
>
> Signed-off-by: Martin Peres
> ---
> src/mapi/glapi/gen/ARB_direct_state_access.xml | 7 +++
> src/mesa/main/test
v2: Review from Laura Ekstrand
- use the transform feedback object lookup wrapper
Signed-off-by: Martin Peres
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 7 ++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/transformfeedback.c | 32 +++
The corresponding piglit tests have been submited to the piglit ML.
Warning: those tests will only work when forcing mesa to expose the DSA
extension.
Martin Peres (6):
main: Added entry point for glCreateTransformFeedbacks
main: Added entry point for glTransformFeedbackBufferBase
main: Ad
v2: Review from Laura Ekstrand
- give more helpful error messages
- factor the lookup code for the xfb and objBuf
- replace some already-existing tabs with spaces
- add comments to explain the cases where xfb == 0 or buffer == 0
- fix the condition for binding the transform buffer or not
Signed-of
v2: review from Laura Ekstrand
- use the refactored code to lookup the objects
- improve some error messages
- factor out the gl method name computation
- better handle the spec differences between the DSA and non-DSA cases
- quote the spec a little more
Signed-off-by: Martin Peres
---
src/mapi/
v2: Review from Laura Ekstrand
- use the transform feedback object lookup wrapper
Signed-off-by: Martin Peres
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 7 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/transformfeedback.c | 29 ++
v2: Review from Laura Ekstrand
- use the transform feedback object lookup wrapper
Signed-off-by: Martin Peres
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 6 ++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/transformfeedback.c | 25 +++
v2: Review from Laura Ekstrand
- generate the name of the gl method once
- shorten some lines to stay in the 78 chars limit
Signed-off-by: Martin Peres
---
src/mapi/glapi/gen/ARB_direct_state_access.xml | 7 +++
src/mesa/main/tests/dispatch_sanity.cpp| 1 +
src/mesa/main/transformfeedb
On Wednesday, January 28, 2015 04:27:23 PM Jason Ekstrand wrote:
> On Wed, Jan 28, 2015 at 4:02 PM, Matt Turner wrote:
>
> > On Tue, Jan 27, 2015 at 5:31 PM, Jason Ekstrand
> > wrote:
> > > diff --git a/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> > b/src/mesa/drivers/dri/i965/brw_fs_nir.cpp
> > >
Paul originally had to reverse engineer these formulas based on the
description about how the sampler works. The description here is not
the easiest to follow - especially given that it's from the Sandybridge
era, when the hardware only did 4x multisampling.
Jordan and I recently found another pa
82 matches
Mail list logo
